DISAL glycosyl donors for efficient glycosylations under acidic conditions: Application to solid-phase oligosaccharide synthesis

Article abstract of DOI:10.1039/b105872j

The use of DISAL (methyl dinitrosalicylate) glycosyl donors in efficient Lewis acid-promoted glycosylations is reported. N-Acetyl-D-glucosamine monosaccharide acceptors are successfully glycosylated at O-6 or O-4 using benzyl- and benzoyl-protected DISAL donors in CH₂Cl₂ or nitromethane in the presence of LiClO₄. The resultant disaccharides are isolated in yields ranging from 35 to 93%. Other Lewis acids such as FeCl₃, TMSOTf, or BF₃·Et₂O also prove efficient for glycosylation of the secondary alcohol cyclohexanol. However, for the synthesis of disaccharides, the mild activation by LiClO₄ gives higher yields. This approach is extended to efficient solid-phase glycosylation of a D-glucosamine derivative anchored by the 2-amino group through a Backbone Amide Linker (BAL) to a polystyrene support.

Full text of DOI:10.1039/b105872j

View Article Online / Journal Homepage / Table of Contents for this issue
DISAL glycosyl donors for efficient glycosylations under acidic
conditions: Application to solid-phase oligosaccharide synthesis1,2
Lars Petersen and Knud J. Jensen*
1
Department of Chemistry, Building 201, Kemitorvet, Technical University of Denmark,
DK-2800 Kgs. Lyngby, Denmark. E-mail: kj@kemi.dtu.dk; Fax: (ϩ45) 45 93 39 68;
Tel: (ϩ45) 45252141
Received (in Cambridge, UK) 3rd July 2001, Accepted 27th July 2001
First published as an Advance Article on the web 3rd September 2001
The use of DISAL (methyl dinitrosalicylate) glycosyl donors in efficient Lewis acid-promoted glycosylations
is reported. N-Acetyl--glucosamine monosaccharide acceptors are successfully glycosylated at O-6 or O-4
using benzyl- and benzoyl-protected DISAL donors in CH2Cl2 or nitromethane in the presence of LiClO4. The
resultant disaccharides are isolated in yields ranging from 35 to 93%. Other Lewis acids such as FeCl3, TMSOTf, or
BF3ؒEt2O also prove efficient for glycosylation of the secondary alcohol cyclohexanol. However, for the synthesis of
disaccharides, the mild activation by LiClO4 gives higher yields. This approach is extended to efficient solid-phase
glycosylation of a -glucosamine derivative anchored by the 2-amino group through a Backbone Amide Linker
(BAL) to a polystyrene support.
Introduction
Cell-surface oligosaccharides are involved in many biological
recognition processes including leukocyte–endothelial cell
adhesion (leukocyte recruitment in inflamed tissue), bacterial
and viral infection, and immunological recognition of tumor
cells and foreign tissue in xenotransplantation.3 Also, free
oligosaccharides can have biological effects such as the rhizo-
bial lipochitin oligosaccharides, which function as nodulation
factors.4 Amino sugars are found in many biologically import-
ant poly- and oligosaccharides, e.g. in glycans of O- and
N-glycoproteins and peptides, lipochitin nodulation factors,
and amino glycoside antibiotics such as streptomycin.
Fig. 1 Generalised structure of DISAL glycosyl donors.
Established methods for glycosylation of aliphatic alcohols,
most notably glycosyl bromides, fluorides, trichloroacetim-
idates, sulfoxides, and n-pentenyls as well as thioglycosides and
glycals, all require activation by a Lewis acid [e.g., AgOTf, BF3ؒ
Et2O, TMSOTf, or dimethyl(methylthio)sulfonium trifluoro-
methanesulfonate (DMTST)] prior to the actual glycosylation
reaction. Waldmann and co-workers have reported activation
of glycosyl fluorides, trichloroacetimidates, and phosphites
using the mild Lewis acid LiClO4.5 This has been ascribed to the
ability of solutions of LiClO4 in Et2O to stabilise polar or ionic
transition states.5 Lubineau and Drouillat have reported LiOTf
as an alternative to LiClO4 as a promoter in glycosylation
reactions and ascribed its effect to general acid catalysis.6 In a
previous paper we have described an efficient method for glyco-
sylation under strictly neutral or mild basic conditions.7 In this
glycosylation technique, the anomeric leaving group on benzyl-
protected donors is methyl 3,5-dinitrosalicylate8 (DISAL) 1 or
its para regioisomer 2 (Fig. 1).
These aryl glycosides were prepared by convenient nucleo-
philic aromatic substitution of an activated aryl fluoride and
were stable on prolonged storage at 5 ЊC and for days in CH2Cl2
and other non-polar solvents. They become efficient glycosyl
donors in polar, aprotic solvents, such as N-methylpyrrolidin-
2-one (NMP), in the absence of Lewis acids. Methanol was
glycosylated stereospecifically and less reactive alcohols, i.e.
monosaccharides, were glycosylated with α-selectivity. This
concept has also been extended to intramolecular glycosyl-
ations.9 Whereas benzyl-protected aryl glycosides were efficient
glycosyl donors under neutral conditions, the analogous
benzoyl-protected donors did not give the expected glycosides,
in part due to trapping of intermediates as the orthoesters.
Solid-phase synthesis has been tremendously successful for
the synthesis of peptides and oligonucleotides, and even small
proteins, due to its inherent simplicity and automatability.10
Interest in the solid-phase synthesis of small organic molecules
has escalated over the past decade.11 When solid-phase syn-
theses are carried out combinatorially, large numbers of struc-
turally diverse compounds can be accessed.12 The most notable
failure of solid-phase synthesis of biopolymers is in the field of
oligosaccharide synthesis. Despite numerous efforts, there is
still no truly general strategy for the solid-phase synthesis of
oligosaccharides.13–15
Here we report on the use of both benzyl- and benzoyl-
protected DISAL glycosyl donors in the presence of ‘classical’
Lewis acid glycosylation promoters, such as BF3ؒEt2O,
TMSOTf, and FeCl3.16 Furthermore, the use of LiClO4 as a
mild activator is reported as an efficient and preferred altern-
ative to these Lewis acids. Finally, DISAL glycosyl donors
were applied to solid-phase oligosaccharide synthesis using a
Backbone Amide Linker (BAL) strategy for anchorage of
-glucosamine derivatives.
Results and discussion
Model glycosylation of cyclohexanol
Initial studies focused on glycosylation of the secondary
alcohol cyclohexanol as a model for carbohydrate acceptors.
DOI: 10.1039/b105872j
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182
This journal is © The Royal Society of Chemistry 2001
2175

View Article Online
Table 1 Model glycosylation of cyclohexanol using benzyl-protected donor 1
Activator
(equiv.)
Time
(t/h)
Temp.
(θ/ЊC)
Yield of 4
(%) (α/β)b
Entry
Donora
Solvent
1
2
3
4
5
6
7
1
1
1
1
1
1
1
BF3ؒEt2O (2.0)
TMSOTf (1.1)
FeCl3 (2.0)
LiClO4 (2.5)
LiClO4 (2.5)
LiClO4–Bu4NI (2.5)
Bu4NI (2.0)
Toluene
Toluene
CH2Cl2
CH2Cl2
Et2O
CH2Cl2
CH3NO2
17
17
2
17
17
17
17
rt
rt
rt
rt
rt
rt
50
97 (2.1 : 1)
87 (2.2 : 1)
89 (1.5 : 1)
93 (1 : 1.1)
84 (2.2 : 1)
81 (6.4 : 1)
90 (4.3 : 1)
a α/β 4 : 1. b Determined by RP-HPLC (215 nm).
Next, we turned to glycosylation of cyclohexanol with
benzoyl-protected donor 3 which had proven inefficient under
neutral conditions. Using up to 10 equiv. of BF3ؒEt2O or
TMSOTf at room temperature left most of the donor
unchanged after 17 h (data not shown). However, raising the
temperature to 40 ЊC in the presence of TMSOTf in 1,2-
dichloroethane, toluene or CH3NO2 resulted in complete con-
sumption of the donor and formation of 5 in 82, 81, and 71%
yield, respectively (Table 2, entries 1–3). The two anomers of
donor 3 had significantly different reactivities in the glycosyl-
ation of cyclohexanol in the presence of BF3ؒEt2O. Whereas
3ꢀ gave incomplete conversion (data not shown), donor 3ꢁ
gave 86% of 5 after 17 h with complete consumption of the
donor (Entry 4). In the case of FeCl3, in the presence of 4 Å
molecular sieves, no reaction occurred for 3ꢀ whereas 3ꢁ gave
the pure β-product in 67% yield (Entries 5 and 6). Again,
LiClO4 proved mild and efficient in activating 3ꢁ in CH3NO2.
After 17 h at 40 ЊC, 5ꢁ was obtained in 81% yield and no 5ꢀ
could be detected (Entry 7).
In some of the above experiments, the kinetically favored
β-glycoside, which was formed initially, under longer reac-
tion times converted into the thermodynamically favored
α-glycoside, most likely due to strong Lewis acid-promoted
transglycosylation in the presence of excess of acceptor,
cyclohexanol (Entries 1–4).
Thus, in the initial screening of Lewis acid-promoted glyco-
sylations with DISAL donors, BF3ؒEt2O and LiClO4 proved to
be the most efficient promoters for glycosylation of cyclohexa-
nol with donors 1 and 3. Best results for 3 were provided by the
β-donor, as the α-anomer was less reactive. Consequently, 3ꢁ
was used in the following glycosylations, in addition to 1.
Fig. 2 Structure of model cyclohexyl glycoside products.
Cyclohexyl glycosides produced (Fig. 2) were readily separated
by analytical RP-HPLC, thus providing information on the
reaction yield and α/β ratio. To assess the efficiency of
the donor, cyclohexanol was used in excess (5 equiv.). Early
experiments pointed to either CH2Cl2 or toluene as reaction
medium. (CH2Cl)2 was used as a higher-boiling equivalent of
CH2Cl2, and CH3CN or CH3NO2 as polar solvents.
Treatment of benzyl-protected donor 1 with cyclohexanol at
room temperature in the presence of BF3ؒEt2O revealed that at
least stoichiometric amounts of Lewis acid were required for
the glycosylation to proceed. Thus, for the reaction in toluene,
sub-stoichiometric amounts of BF3ؒEt2O left much of the
donor unchanged after 17 h, 1.1 equiv. left only a trace of
donor, and 2 equiv. produced cyclohexyl glycoside 4 in 97%
yield (Table 1, Entry 1). Interestingly, the α-anomer was
consumed first in this reaction. TMSOTf proved to be more
activating, as only a trace of donor was observed with 0.5 equiv.
after 17 h and total conversion was achieved with 1.1 equiv. to
yield 87% of 4 (Entry 2). FeCl3, solubilised in CH2Cl2, also
proved efficient and 2 equiv. gave total conversion to 89% of 4
after only 2 h (Entry 3).
The requirement for more than stoichiometric amounts
of BF3ؒEt2O deserves a comment. It has been reported that
Glycosylation of carbohydrate acceptors
4-methoxysalicylaldehyde readily forms
a tight complex
with BF3, releasing one equivalent of HF.17 We speculated
that the released phenoxide anion of methyl 2-hydroxy-3,5-
dinitrobenzoate was not only acting as a general Lewis
base towards Lewis acids, but also reacted with BF3 in a
specific manner similar to its reaction with 4-methoxy-
salicylaldehyde.
With these optimised conditions in hand, we turned to
glycosylation of monosaccharide acceptors, first with benzyl-
protected donor 1ꢀ,ꢁ. We initially used diisopropylidene-
protected acceptors 6 and 7 (Scheme 1), which could be glyco-
sylated by DISAL donors under neutral conditions.7 However,
they proved not to be sufficiently stable in the presence of BF3ؒ
Et2O. Turning to the more stable benzyl-protected primary and
secondary acceptors 819 and 9,20 respectively, moderate yields
of the 1,6- and 1,4-linked disaccharides 10 and 11, respectively,
were obtained (Table 3, Entries 1 and 2). However, lowering the
temperature to 0 ЊC and shortening the reaction time gave rise
to a significantly improved yield of 82% of 10 (Entry 3). Lower-
ing the temperature further to Ϫ44 ЊC slowed the reaction
significantly and at Ϫ78 ЊC no glycosylation was observed (data
not shown). Also, substituting BF3ؒEt2O by either FeCl3 or
TMSOTf in the glycosylation with 1 in CH2Cl2 gave rise to
complex reaction mixtures (data not shown).
It was rewarding to see that the very mild Lewis acid LiClO4
also proved efficient in activating donor 1. Even though only
sparingly soluble in CH2Cl2, LiClO4 readily gave a near quanti-
tative yield of 4 as an α/β mixture (Entry 4). The effect of the
solvent on anomeric selectivity was studied. Incomplete reac-
tion was observed in CH3CN (data not shown). However, full
conversion and slightly higher α-selectivity was observed in
Et2O (Entry 5). Addition of an auxiliary nucleophile, Bu4NI, to
the glycosylation in CH2Cl2 in the presence of LiClO4 gave both
a good yield and high α-selectivity (Entry 6). An NMR experi-
ment with donor 1, Bu4NI, and LiClO4 (2 equiv. each) in CDCl3
at room temperature showed a doublet at δ 6.84 (J 3.8 Hz)18
characteristic for the α-glucosyl iodide after only 30 min (≈ 20%
conversion). With Bu4NI alone, good yield and α-selectivity
was achieved but it required slightly more forcing conditions
(Entry 7).
The apparent breakdown of initially formed disaccharide in
the presence of BF3ؒEt2O in extended reactions at rt is note-
worthy, as glycosides generally are considered stable to BF3ؒ
Et2O. We assumed that the HF generated by formation of
the above-mentioned complex between BF3ؒEt2O and released
2176
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182

View Article Online
Table 2 Model glycosylation of cyclohexanol using benzoyl-protected donor 3
Entry
Donor
Activator (equiv.)
Solvent
Time (t/h)
Temp. (θ/ЊC)
Yield of 5 (%) (α/β)a
1
2
3
4
5
6
7
3
TMSOTf (10)
TMSOTf (10)
TMSOTf (10)
BF3ؒEt2O (5.0)
FeCl3 (5.0)b
(CH2Cl)2
Toluene
CH3NO2
(CH2Cl)2
CH3CN
CH3CN
CH3NO2
17
17
17
17
17
17
17
40
40
40
40
40
40
40
82 (7.2 : 1)
81 (6.7 : 1)
71 (7.2 : 1)
3
3
3ꢁ
3ꢀ
3ꢁ
3ꢁ
86 (9.6 : 1)
c
FeCl3 (5.0)b
67 (β only)
81 (β only)
LiClO4 (2.5)
a Determined by RP-HPLC (215 nm). b 4 Å molecular sieves. c No reaction.
Table 3 Synthesis of disaccharides
Entry
Donora
Acceptor
Activator (equiv.)
Solvent
Time (t/h)
Temp. (θ/ЊC)
Product
Yield (%) (α/β)b
1
2
3
4
5
6
7
8
9
1ꢀ,ꢁ
1ꢀ,ꢁ
1ꢀ,ꢁ
3ꢁ
1ꢀ,ꢁ
1ꢀ,ꢁ
3ꢁe
8
9
BF3ؒEt2O (1.8)
BF3ؒEt2O (1.8)
BF3ؒEt2O (1.8)
BF3ؒEt2O (1.8)c
BF3ؒEt2O (1.8)c
LiClO4 (2.0)c, d
LiClO4 (3.7)
CH2Cl2
CH2Cl2
CH2Cl2
toluene
CH2Cl2
CH2Cl2
CH3NO2
CH2Cl2
CH3NO2
17
17
0.5
17
17
1.5
17
5
17
rt
rt
0
40
rt
rt
40
rt
40
10
11
10
12
16
16
17
22
23
40 (3.3 : 1)
34 (1 : 3.3)
82 (1 : 1)
46 (β)
77 (1.4 : 1)
93 (1.6 : 1)
91 (β)
8
8
14
14
14
19
19
1ꢀ,ꢁ
3ꢁ
LiClO4 (2.0)c, d
LiClO4 (3.7)
82 (1.9 : 1)
35 (β)
a 1.5 equiv. donor. b Isolated yield. c Molecular sieves 4 Å. d 2.0 equiv. of Li2CO3 added. e 2.0 equiv. donor.
from N-acetyl--glucosamine by Fischer glycosylation, fol-
lowed by tritylation21 of the 6-OH, and benzylation.22 This was
achieved in good yields with no chromatographic steps. Select-
ive hydrolysis of the trityl ether gave primary acceptor 14.
Furthermore, hydrolysis of 13 under more forcing conditions
gave amine 15 in good yield. Glycosylation of 14 in CH2Cl2 at
room temperature with benzyl donor 1 activated by BF3ؒEt2O
yielded 77% of disaccharide 16ꢀ,ꢁ (Scheme 2; Table 3, Entry 5).
A significantly improved yield was obtained by activation with
LiClO4 and product 16α,β was isolated in 93% yield, again as an
anomeric mixture (Entry 6). Note at this point that addition of
solid Li2CO3 as an acid scavenger was tolerated by the protocol.
This quenched or slowed down the activation when using BF3ؒ
Et2O. Acceptor 14 was also glycosylated successfully with
donor 3ꢁ and the pure β-product 17ꢁ was obtained in an
excellent 91% yield (Entry 7).
To obtain the corresponding acceptor with a free O-4, the
known benzylidene 18 was selectively ring-opened20 using
triethylsilane and BF3ؒEt2O, and the desired product 19 was
isolated in 74% yield. Furthermore, benzylidene 18 was
selectively N-deacetylated in good yield (84%) with Tf2O and
pyridine following the procedure by Charette and Chua.23 This
procedure was originally developed for the conversion of
mono- and disubstituted amides into esters and free amines,
but was readily adapted for N-deacetylation of GlcNAc deriv-
atives. Resultant benzylidene 20 was subsequently ring-opened
by prolonged treatment with triethylsilane and BF3ؒEt2O and a
moderate yield (55%) of amine 21 was isolated. The 1
4
linkage to GlcNAc acceptors is among the most difficult to
establish. It was rewarding to see that secondary acceptor 19
was glycosylated efficiently using an excess of LiClO4 in
CH3NO2. Thus, using benzyl donor 1ꢀ,ꢁ, disaccharide 22α,β
was isolated as an anomeric mixture in 82% yield (Entry 8). For
the benzoyl donor, disaccharide 23 was obtained as the pure
β-anomer in a moderate isolated yield (Entry 9).
Scheme 1
phenoxide caused the decomposition of formed products. Sev-
eral scavengers were tested to quench formed HF, e.g. silicon-
containing compounds and sterically hindered bases (data not
shown). We found that addition of molecular sieves suppressed
the decomposition significantly and allowed for glycosylation at
higher temperatures and longer reaction times. Thus, primary
acceptor 8 was successfully glycosylated by benzoyl-protected
donor 3ꢁ at 40 ЊC in toluene in the presence of BF3ؒEt2O to give
β-glycoside 12ꢁ in 46% yield (Entry 4).
Solid-phase glycosylations
For solid-phase glycosylations, we relied on a BAL strategy24
for anchoring of -glucosamine derivatives. This allows for
safety-catch attachment through the amine, thereby providing
high stability towards Lewis acids during glycosylation and
Next, we turned to the glycosylation of -glucosamine
derivatives. Intermediate 13 (Scheme 2) was synthesized starting
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182
2177

View Article Online
Table 4 Solid-phase glycosylation of polystyrene-bound acceptors 25 and 26
Entry
Donor (equiv.)
Resin
Activator (equiv.)
Solvent
Temp. (θ/ЊC)
Product
Yield (%)a
Recovery (%)b
1
2
3
4
5
6
7
8
9
1 (3.0)
25
25
25
25
25
25
25
26
25
26
25
BF3ؒEt2O (3.6)
BF3ؒEt2O (6.0)
BF3ؒEt2O (12.0)
BF3ؒEt2O (3.6)
BF3ؒEt2O (6.0)
BF3ؒEt2O (12.0)
LiClO4 (25)
LiClO4 (25)
LiClO4 (38)
LiClO4 (38)
Bu4NI (30)
CH2Cl2
CH2Cl2
CH2Cl2
Toluene
Toluene
Toluene
CH2Cl2
CH2Cl2
CH3NO2
CH3NO2
CH3NO2
rt
rt
rt
40
40
40
rt
16
16
16
17
17
17
16
22
17
23
16
12
25
80
1
17
75
98
52
88
7
97
86
79
98
90
83
65
79
83
97
22
1 (5.0)
1 (10.0)
3 (3.0)
3 (5.0)
3 (10.0)
1 (10.0)
1 (10.0)
3 (15.0)
3 (15.0)
1 (10.0)
rt
40
40
50
10
11
84
a Ratio of product to the sum of unchanged acceptor and product calculated from standard curves on RP-HPLC (215 nm). b Sum of unchanged
acceptor (27 or 28) and disaccharide (16, 17, 22, or 23) compared with initially measured loading.
Scheme 2 Reagents and conditions: (i) THF–3 M aq. HCl (5 : 1), 50 ЊC, 17 h; (ii) THF–3 M aq. HCl (2 : 1), reflux, 3 d; (iii) Et3SiH, BF3ؒEt2O
(2 equiv.), CH2Cl2, 0 ЊC; (iv) Tf2O, pyridine, Ϫ40 ЊC; then excess of EtOH; (v) Et3SiH, BF3ؒEt2O (3 equiv.), CH2Cl2, 0 ЊC.
efficient release of final products by TFA after acetylation of
the nitrogen.
of acetylated, secondary acceptor 28 gave a loading of 0.27
mmol gϪ1 (84%, 5 steps). Acetylated acceptors 27 and 28 were
both identified by NMR and MS. It should be noted that add-
ition of DMAP to the acetylation mixtures should be avoided
as it gave rise to impurities.
First, 5-(2-formyl-3,5-dimethoxyphenoxy)valeric acid (o-
PALdehyde, 24) was coupled to aminomethylated polystyrene
(PS) or poly(ethylene glycol) polyacrylamide (PEGA) supports,
both with an initial loading of 0.40 mmol gϪ1 using a standard
peptide protocol.14 Next, primary acceptor 15 was anchored by
reductive amination in the presence of NaBH3CN. Reductive
amination in DMF–AcOH (99 : 1)25 proceeded well on a
PS support but gave low yields on PEGA.26 Previously, it was
observed that reductive amination on PALdehyde in solution
gave double alkylation in DMF but monoalkylation in
methanol.24a We reasoned that the apparent low loading was
due to double alkylation of the amino group and attributed this
to reaction with two PALdehyde moieties in flexible PEGA
supports due to lack of efficient ‘site isolation’.26 Thus, when
the reductive amination on PEGA was carried out in methanol,
good yields were obtained. After acetylation, monosaccharide
product 27 was released with TFA–water (19 : 1) (Scheme 3),
which gave a loading of 0.30 mmol gϪ1 for PS (94%, 5 steps) and
0.028 mmol gϪ1 for PEGA (80%, 5 steps); the latter swelled in
NMP. PEGA proved difficult to handle in the dried state, and
therefore was handled swelled in dry NMP. Loading was
calculated from drained weight, i.e. swelled in NMP, thus
giving a seemingly low loading. Reductive amination with
amine 21 on PS-supported o-PALdehyde 24 followed by release
Initial efforts on solid-phase glycosylation with DISAL
donors started from the previously reported neutral conditions
in solution.7 However, the benzyl-protected donor 1ꢀ,ꢁ (up to
10 equiv.) in NMP at 40–80 ЊC glycosylated PS- and PEGA-
bound acceptor in only 0–9% yield. Next, we used Lewis acid-
promoted glycosylations as above. Promoters TMSOTf and
FeCl3 proved unsuitable for glycosylations with donors 1ꢀ,ꢁ or
3ꢁ on the solid phase. However, with donor 1ꢀ,ꢁ in the presence
of BF3ؒEt2O and 4 Å molecular sieves, efficient glycosylation of
polystyrene-bound acceptor 25 was achieved. Under the same
conditions, the PEGA-bound acceptor was not glycosylated,
presumably due to quenching of the Lewis acid by the resin.
Glycosylations on PS, on the other hand, gave disaccharide
16ꢀ,ꢁ (α/β ≈ 2 : 1) and monosaccharide 27 in ratios of 12 : 88, 25
: 75 and 80 : 20 for a glycosylation using 3, 5, or 10 equiv. of
1ꢀ,ꢁ, respectively (Table 4, Entries 1–3). Substituting glycosyl
donor 1ꢀ,ꢁ by the more reactive 2ꢀ,ꢁ did not increase the yield
(data not shown). Benzoyl-protected donor 3ꢁ in toluene at
40 ЊC in the presence of BF3ؒEt2O yielded ratios of the expected
disaccharide 17ꢁ to monosaccharide 27 of 1 : 99, 17 : 83 and
75 : 25, again with 3, 5, or 10 equiv. of donor, respectively
2178
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182

View Article Online
Scheme 3 Reagents and conditions: (i) HOBt (2 equiv.), HBTU (1.9 equiv.), DIEA (3.9 equiv.), DMF, rt, 16 h; (ii) CH2Cl2–Ac2O–pyridine (18 : 1 : 1),
16 h; (iii) 15 or 21 (2 equiv.), NaBH3CN (10 equiv.), DMF–AcOH (99 : 1) (PS) or MeOH–AcOH (99 : 1) (PEGA); (iv) Ac2O–pyridine (2 : 1), 16 h;
(v) TFA–water (19 : 1), 60 min; (vi) glycosylation: see Table 4.
(Entries 4–6). Thus, 3ꢁ also proved to be efficient in glycosyl-
ation reactions on the solid phase.
competent promoter. Benzyl-protected DISAL donors gave
predominantly the α-glycosides, whereas the benzoyl-protected
donor in general gave β-glycosides. The preferred conditions for
glycosylations at O-6 and O-4 were 1.5 equiv. of donor in a
solution of LiClO4 at rt (CH2Cl2) or 40 ЊC (CH3NO2). Using
this protocol, disaccharides were prepared in up to 93 and 82%
yield for the (1 6)- and (1 4)-linkage, respectively. Finally,
the optimised conditions were adapted to solid-phase synthesis
of 1 6- and 1 4-linked disaccharides in good to moderate
yields.
We then turned to LiClO4 activation of glycosyl donors
in solid-phase glycosylations. Donor 1 (10 equiv.) again
proved efficient and gave an almost quantitative conversion to
disaccharide 16 (Entry 7), however, with a lower recovery
than for BF3ؒEt2O (Entry 3). The very hindered resin-bound
acceptor 26 was also glycosylated under these conditions to give
a 52% yield of disaccharide 22 (Entry 8). Donor 3ꢁ with activ-
ation by LiClO4 also glycosylated both resin-bound acceptors
25 and 26 to give disaccharides 17 and 23 in 88% (α/β 10 : 1;
Entry 9) and 7% yield, respectively (Entry 10). The conditions
under which cyclohexanol was efficiently glycosylated using
donor 1 together with Bu4NI (Table 1; Entry 7) were also tested
in solid-phase glycosylation of 25. A high conversion was
achieved together with a high α-selectivity (α/β 13 : 1, Entry 11).
However, a significant amount of material appeared to be lost
from the resin under these conditions.
Experimental
General
Mps were measured on a Danotherm melting point apparatus
and are uncorrected. All solvents were distilled and/or stored
1
over 3 Å or 4 Å molecular sieves as appropriate. H-NMR
spectra were recorded on either a Varian Mercury 300 operating
at 300.06 MHz equipped with a 4-nuclei probe or a Varian
Unity Inova 500 operating at 499.87 MHz equipped with a
z- (single axis) PFG inverse detection C-H-P probe. 13C-NMR
spectra were recorded on a Varian Mercury 300 operating at
75.46 MHz. Chemical shift (δ)-values were in ppm and coupling
constants (J) in Hz. All assignments were supported by 2D
homonuclear chemical-shift correlation spectroscopy (gCOSY)
and heteronuclear single quantum correlated spectroscopy
(gHSQC) experiments. Thin-layer chromatography (TLC) was
performed on Merck Silica Gel 60 F254 plates and spots were
visualised by UV light at 254 nm and/or spraying with 10% aq.
H2SO4 followed by heating. Molecular sieves (4 Å) for glyco-
sylation reactions were crushed under argon followed by acti-
vation at 150 ЊC under high vacuum for 16 h. Vacuum liquid
Thus, a high degree of glycosylation of O-6 of a resin-bound
glucosamine acceptor was achieved using 10–15 equiv. of
donors in the presence of a large excess of the very mild Lewis
acid LiClO4 in solutions of CH3NO2 or CH2Cl2 as solvent.
Conclusions
We have described the activation of DISAL glycosyl donors by
Lewis acids to promote glycosylation reactions in non-polar
solvents. Whereas benzoyl-protected DISAL donors previously
had proven inefficient for glycosylation under neutral condi-
tions, Lewis acids activated these ‘disarmed’ donors efficiently.
While some conventional promoters of glycosylations, such as
BF3ؒEt2O and TMSOTf, also proved their efficiency here, it was
remarkable that the very mild Lewis acid LiClO4 also was a
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182
2179

View Article Online
chromatography (VLC) was carried out on Merck Silica Gel
60H. HPLC analyses were conducted with a Waters system [600
control unit, 996 photodiode array (PDA) detector, 717 Plus
autosampler, Millenium32 control software] on a Waters Nova-
Pak or XTerra C18 column (3.9 × 5.0 mm cartridge; 4 µm
particle size) using a linear gradient of 0.1% aq. TFA (A) and
0.1% TFA in CH3CN (B): 0 min: 0% B, 2 min: 0% B, 5 min: 50%
B, 12 min: 95% B, 13 min: 95% B, 13.5 min: 0% B, 20 min: 0% B.
Monitoring was from 200 to 400 nm, integrations were per-
formed at 215 and 265 nm, and individual peaks were analyzed
by their UV spectra. The purity of compounds was determined
from integrations at 215 nm. MS analyses (electrospray,
positive mode) were performed on a Micromass LCT mass
spectrometer.
added (0.6–1.0 cm3) followed by a pre-activation period (30–60
min) for removal of residual water by the molecular sieves.
Lewis acid as indicated was either diluted in the solvent indi-
cated and added with a Hamilton syringe (BF3ؒEt2O) or added
as solid (LiClO4). Reactions that required heating were placed
on an Eppendorf Thermomixer 5436 (combined heater–shaker)
or on a well-equilibrated sand-bath. After agitation for 24 h
(500 minϪ1), the resin was washed successively with DMF (3 × 2
cm3), CH2Cl2–MeOH (2 : 1; 3 × 2 cm3), and CH2Cl2 (5 × 2 cm3).
Acetic anhydride (600 mm3) and pyridine (300 mm3) were
added, and after shaking for 16 h the resin was washed as
before. Acetylated product and remaining acceptor was
released from the resin using TFA–water (19; 1, 0.5 cm3) within
60 min. The cleavage mixture and washes with CH2Cl2 (5 × 2
cm3) were collected and concentrated. The residue was dis-
solved in CH3CN, diluted to 5.00 cm3, membrane filtered (0.45
µm), and a sample injected into analytical HPLC. Cleaved
amounts were quantified by integration from HPLC and
comparison with standard curves (analytical standards of 2,3,4,
6-tetra-O-benzyl--glucose and 1,2,3,4,6-penta-O-benzoyl--
glucopyranose). Loadings of resin-bound acceptors were
determined in the same fashion, but without the glycosylation
step.
General method for model glycosylations and HPLC analysis
(Tables 1 and 2)
The initial evaluation of solvents for the glycosylation was per-
formed by treating glycosyl donor 1ꢀ,ꢁ or 3ꢀ,ꢁ (typically 0.02
mmol) with cyclohexanol and an activator in the solvent indi-
cated (400–600 mm3) on an Eppendorf Thermomixer 5436
(combined heater/shaker). After shaking and heating for the
period indicated, a sample (10–40 mm3) was diluted with
CH3CN (0.70–1.00 cm3) and analyzed by analytical RP-HPLC.
Reported yields were based on integrated areas at 215 nm.
Synthesis of glycosyl donors 1ꢀ,ꢁ, 2ꢀ,ꢁ, and 3ꢀ,ꢁ
Glycosyl donors were synthesised as previously described.7
Anomeric ratios in donors used: 1 (α/β 4.0 : 1), 2 (α/β 4.1 : 1), 3
(α/β 1 : 22.3).
General method for glycosylations and isolation (Table 3)
Glycosylation using 1ꢀ,ꢁ. Glycosyl donor (1.5 equiv.) and
acceptor (8, 9, 14, or 19; 0.1 mmol) were weighed off in a 5 cm3
polypropylene test tube, mounted with a septum and needle,
evacuated in high vacuum and filled with argon, repeated twice,
and then dried under high vacuum for 1–2 h. 50–100 mg of 4 Å
molecular sieves were added followed by addition of the solvent
indicated (1 cm3). The mixture was now agitated overnight to
remove residual water present. The activator was added as a
solution of BF3ؒEt2O in the solvent or as solid LiClO4.
Cyclohexyl 2,3,4,6-tetra-O-benzyl-ꢀ,ꢁ-D-glucopyranoside 4ꢀ,ꢁ
Observed NMR data were identical with literature values.27
Cyclohexyl 2,3,4,6-tetra-O-benzoyl-ꢀ-D-glucopyranoside 5ꢀ
Mp 91–92 ЊC; δH (300 MHz; CDCl3) 8.04–7.82 (8 H, m), 7.60–
7.22 (12 H, m), 6.19 (1 H, t, J 9.8, 3-H), 5.63 (1 H, t, J 9.7, 4-H),
5.50 (1 H, d, J1,2 3.9, 1-H), 5.26 (1 H, dd, J 2,3 10.1 and J1,2 3.9,
2-H), 4.62–4.42 (3 H, m, 5-H and 6-H2), 3.62 (1 H, m), 2.0–1.1
(10 H, m); δC (75 MHz; CDCl3) 166.4, 166.0 (2 × C), 165.6,
134–127 (m), 94.9 (1α-C), 77.8, 72.4, 70.9, 70.0, 68.1, 63.5, 33.7,
31.8, 25.7, 24.2, 23.9; m/z 696.1 ([M ϩ NH4]ϩ. C40H42NO10
requires m/z, 696.3).
Glycosylations using 3ꢁ. Glycosyl donor (1.5–2.0 equiv.) and
acceptor (8, 14, or 19; 0.1 mmol) were weighed off in 1.5 cm3
Eppendorf centrifuge tubes together with LiClO4, if required.
Solvent indicated was added (1 cm3), followed by addition of a
solution of BF3ؒEt2O in the solvent, if required. The mixture
was then agitated and heated on an Eppendorf Thermomixer
5436 (combined heater/shaker).
Cyclohexyl 2,3,4,6-tetra-O-benzoyl-ꢁ-D-glucopyranoside 5ꢁ
Mp 97.5–100 ЊC; δH (300 MHz, CDCl3) 8.04 (8 H, m), 7.56
(12 H, m), 5.90 (1 H, t, J 9.5, 3-H), 5.64 (1 H, t, J 9.8, 4-H), 5.50
(1 H, dd, J2,3 9.8 and J1,2 7.9, 2-H), 4.94 (1 H, d, J1,2 7.8, 1-H),
4.62 (1 H, dd, Jgem 12.1 and J5,6 3.4, 6-H), 4.52 (1 H, dd, Jgem
12.2 and J5,6 6.0, 6-HЈ), 4.19–4.10 (1 H, m, 5-H), 3.72–3.62 (1 H,
m), 2.0–1.1 (10H, m); δC (75 MHz; CDCl3) 166.3, 166.0, 165.4,
165.2, 134–127 (m), 100.1 (1β-C), 78.7, 73.3, 72.4, 72.3, 70.4,
63.7, 33.5, 31.9, 25.6, 23.9, 23.8; m/z 696.1 ([M ϩ NH4]ϩ.
C40H42NO10 requires m/z, 696.3).
Purification. For reactions in CH2Cl2 or toluene the reaction
mixture was diluted with CH2Cl2 (40 cm3), washed successively
with 1 M aq. HCl (2 × 20 cm3) and 0.5 M aq. NaOH (2 × 20
cm3), dried (MgSO4) and concentrated. The residue was dis-
solved in CH3CN, membrane filtered (0.45 µm) using CH3CN,
and products were isolated by preparative HPLC using a gradi-
ent of CH3CN in water. For reactions in nitromethane, the reac-
tion mixture was membrane filtered (0.45 µm) directly using
CH3CN without extraction prior to preparative HPLC.
Appropriate fractions were concentrated in vacuo at 30 ЊC,
traces of water were removed by co-evaporation once each with
CH3CN and CH2Cl2, and the product was dried under high
vacuum to give the isolated yields stated.
Methyl 2,3,4-tri-O-benzyl-6-O-(2,3,4,6-tetra-O-benzyl-ꢀ,ꢁ-D-
glucopyranosyl)-ꢀ-D-glucopyranoside 10ꢀ,ꢁ
Isolated yield after preparative HPLC: 82% (α/β ≈ 1 : 1).
Observed NMR data were identical with literature values.27
General method for solid-phase glycosylations and determination
of loadings (Table 4)
Methyl 2,3,6-tri-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-ꢀ,ꢁ-D-
glucopyranosyl)-ꢀ-D-glucopyranoside 11ꢀ,ꢁ
The resin (PS, 15–20 mg, ≈4–6 µmol acceptor), glycosyl donor
(1ꢀ,ꢁ or 3ꢁ), and 4 Å molecular sieves (≈50 mg) were weighed
off in a 2 cm3 disposable syringe fitted with a polypropylene
filter and a Teflon valve. The syringes were mounted with septa
and evacuated in high vacuum and then filled with argon. The
procedure was repeated twice, followed by drying under high
vacuum for 1–2 h. Upon inlet of argon, the valves were closed
or if heating was required exchanged for a stopper, and solvent
Isolated yield after preparative HPLC: 34% (α/β 1 : 3.3).
Observed NMR data were identical with literature values.27
Methyl 2,3,4-tri-O-benzyl-6-O-(2,3,4,6-tetra-O-benzoyl-ꢁ-D-
glucopyranosyl)-ꢀ-D-glucopyranoside 12
Isolated yield after preparative HPLC: 46% (pure β). Observed
NMR data were identical with literature values.27
2180
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182

View Article Online
Methyl 2-acetamido-3,4-di-O-benzyl-2-deoxy-ꢀ-D-gluco-
[lit.,29 144–145 (from acetone–Et2O–hexane)]; purity > 99%
pyranoside28 14
(HPLC); δH (300 MHz; CDCl3) 7.40–7.25 (10 H, m), 5.42 (1 H,
d, J 9.0, NH), 4.73 (2 H, m, PhCH2), 4.68 (1 H, d, J 3.7, 1α-H),
4.59 (2 H, m, PhCH2), 4.22 (1 H, ddd, J 10.3, 9.0 and 3.7, 2-H),
3.80–3.68 (4 H, m, 4-H, 5-H and 6-H2), 3.56 (1 H, dd, J 10.3
and 7.8, 3-H), 3.34 (3 H, s, OMe), 2.82 (1 H, d, J 2.1, OH), 1.90
(3 H, s, NAc); δC (75 MHz; CDCl3) 170.0, 138.7, 138.1, 128.7–
127.8 (m), 98.9, 80.1, 74.0, 73.9, 72.4, 70.6, 70.9, 55.3, 52.1,
23.7.
Trityl ether 13 (658 mg, 1.0 mmol) was dissolved in THF
(5 cm3), 3 M aq. HCl (1.0 cm3) was added, and the reaction
mixture was heated for 17 h at 50 ЊC by means of an oil-bath.
The solution was then diluted with ethyl acetate (40 cm3),
washed once with brine (20 cm3), dried (MgSO4), and concen-
trated in vacuo. The crude product was then subjected to VLC
on silica gel (CH2Cl2–acetone 4 : 1) and fractions containing a
compound with Rf 0.12 (CH2Cl2–acetone 4 : 1) were concen-
trated in vacuo. The resulting gel was lyophilised from pentane
to yield 14 (292 mg, 70%) as a colorless solid, mp 181–182 ЊC
(lit.,28 186–188 ЊC); purity > 99% (HPLC); observed NMR data
were identical with literature values.28
Methyl 2-amino-3-O-benzyl-4,6-O-benzylidene-2-deoxy-ꢀ-D-
glucopyranoside 20
Adapting a procedure by Charette and Chua,23 benzylidene 18
(827 mg, 2.0 mmol) in a dry flask was dissolved in dry CH2Cl2
(30 cm3) under argon. Dry pyridine (0.50 cm3, 6.0 mmol) was
added and the reaction mixture was cooled to Ϫ40 ЊC by means
of a bath of solid CO2 in CH3CN. Tf2O (0.50 cm3, 3.0 mmol)
was added dropwise with a syringe under stirring and the mix-
ture was allowed to warm to 0 ЊC over a period of 2.5 h. After
another 4 h at 0 ЊC, dry EtOH (5 cm3) was added and the
mixture warmed to rt overnight. Additional CH2Cl2 (20 cm3)
was added and the organic phase was washed successively with
0.5 M aq. HCl (2 × 15 cm3) and saturated aq. NaHCO3 (2 × 15
cm3), dried (Na2SO4), and concentrated to a slightly yellow
solid (0.79 g, from Et2O–pentane). The residue was purified by
VLC (CH2Cl2–MeOH 99 : 1) and fractions showing Rf 0.29
(CH2Cl2–MeOH 95 : 5) were concentrated to give 20 (621 mg,
84%) as a colorless solid, mp 104–110 ЊC; δH (300 MHz; CDCl3)
7.55–7.10 (10 H, m), 5.59 (1 H, s), 5.05 (1 H, d, J 11.5), 4.75
(1 H, d, J 3.7, 1α-H), 4.68 (1 H, d, J 11.4), 4.29 (1 H, dd, J 9.5
and 4.1, 6-H), 3.90–3.73 (2 H, m, 5-H and 6-HЈ), 3.70–3.58
(2 H, m, 3- and 4-H), 3.40 (3 H, s, OMe), 2.86 (1 H, dd, J 9.2
and 3.7, 2-H), 1.57 (2 H, br s, NH2); δC (75 MHz; CDCl3) 138.8,
137.8, 129.1–127.9 (m), 126.3, 126.2, 101.5, 101.4, 83.5, 80.2,
75.2, 69.5, 63.1, 56.2, 55.5; m/z 372.17 ([M ϩ H]ϩ. C21H26NO5
requires m/z, 372.18).
Methyl 2-amino-3,4-di-O-benzyl-2-deoxy-ꢀ-D-glucopyranoside
15
Trityl ether 13 (658 mg, 1.0 mmol) was dissolved in THF
(5 cm3), 3 M aq. HCl (2.5 cm3) was added, and the reaction
mixture was heated at reflux for 3 d by means of an oil-bath.
The solution was then diluted with ethyl acetate (40 cm3) and
extracted with 1 M aq. HCl (2 × 20 cm3). The combined aq.
phases were treated with conc. NaOH (28%) until pH > 11 and
extracted with ethyl acetate (2 × 20 cm3). The combined organic
phases were washed with brine (10 cm3), dried (Na2SO4), and
concentrated in vacuo. The crude syrup was then subjected to
VLC on silica gel (CH2Cl2–acetone–MeOH 75 : 20 : 5) and
fractions containing a compound with Rf 0.25 (CH2Cl2–
acetone–MeOH 72.5 : 20 : 7.5) were concentrated in vacuo to
yield amine 15 as a colorless syrup that solidified on storage and
was lyophilised from pentane (226 mg, 61%), mp 68–70 ЊC;
purity > 97% (HPLC); δH (500 MHz; CDCl3) 7.36–7.25 (10 H,
m), 4.98 (1 H, d, Jgem 11.3), 4.87 (1 H, d, Jgem 11.0), 4.75–4.67
(3 H, m, 2 × PhCH2 ϩ 1-H), 3.83 (1H, dd, Jgem 11.5 and J5,6 2.1,
6-H), 3.75 (1 H, dd, Jgem 11.9 and J5,6 3.4, 6-HЈ), 3.71–3.67 (1 H,
m, 5-H), 3.59 (1 H, t, J 9.0, 3-H), 3.54 (1 H, t, J 9.0, 4-H), 3.36
(3 H, s, OMe), 2.77 (1 H, dd, J2,3 9.0 and J1,2 3.4, 2-H); δC (75
MHz; CDCl3) 138.8, 138.3, 129–128 (m), 100.9 (1α-C), 84.1,
78.9, 75.9, 75.0, 71.7, 62.1, 56.4, 55.4; m/z 374.19 ([M ϩ H]ϩ.
C21H28NO5 requires m/z, 374.20).
Methyl 2-amino-3,6-di-O-benzyl-2-deoxy-ꢀ-D-glucopyranoside
21
Benzylidene 20 was ring-opened by the method of Debenham
and Toone20 using triethylsilane and BF3ؒEt2O (3 equiv.) for 16
h. Isolated yield of 21 after VLC (CH2Cl2–MeOH 99 : 1): 55%
(colorless syrup); purity > 98% (HPLC); δH (300 MHz; CDCl3)
7.40–7.25 (10 H, m), 4.92 (1 H, d, J 11.9, PhCH2), 4.77 (1 H, d,
J 11.9, PhCH2), 4.70 (1 H, d, J 3.4, 1α-H), 4.63 (1 H, d, J 12.0,
PhCH2), 4.55 (1 H, d, J 12.0, PhCH2), 3.78–3.61 (4 H, m, 4-H,
5-H and 6Ј-H2), 3.42 (1 H, dd, J 9.9 and 8.4, 3-H), 3.38 (3 H, s,
OMe), 2.77 (1 H, dd, J 9.9 and 3.4, 2-H), ≈1.5 (≈2 H, br s,
NH2); δC (75 MHz; CDCl3) 139.0, 138.1, 128.8–127.9 (m),
101.0, 84.1, 75.5, 74.0, 73.3, 70.8, 70.3, 55.6, 55.4; m/z 374.19
([M ϩ H]ϩ. C21H28NO5 requires m/z, 374.20).
Methyl 2-acetamido-3,4-di-O-benzyl-2-deoxy-6-O-(2,3,4,6-
tetra-O-benzyl-ꢀ,ꢁ-D-glucopyranosyl)-ꢀ-D-glucopyranoside
16ꢀ,ꢁ
Isolated yield after preparative HPLC: Activation by BF3ؒEt2O
(77%, α/β 1.4 : 1) and LiClO4 (93%, α/β 1.6 : 1). The two
anomers proved inseparable by preparative HPLC; purity
> 99% (HPLC); selected NMR-data: δH (300 MHz; CDCl3)
7.42–7.12 (30 H, m, Bn), 5.27 (1 H, 2 d, NH), 3.32 (s, α-OMe)
and 3.29 (s, β-OMe), 1.87 (3 H, s, NAc); δC (75 MHz; CDCl3)
16ꢀ: 104.1 (1Јβ-C), 98.8 (1α-C); 16ꢁ: 97.5 (1Јα-C), 98.6 (1α-C);
m/z (anomeric mixture) 960.43 ([M ϩ Na]ϩ. C57H63NNaO11
requires m/z, 960.43).
Methyl 2-acetamido-3,6-di-O-benzyl-2-deoxy-4-O-(2,3,4,6-
tetra-O-benzyl-ꢀ,ꢁ-D-glucopyranosyl)-ꢀ-D-glucopyranoside
22ꢀ,ꢁ
Methyl 2-acetamido-3,4-di-O-benzyl-2-deoxy-6-O-(2,3,4,6-
tetra-O-benzoyl-ꢁ-D-glucopyranosyl)-ꢀ-D-glucopyranoside 17ꢁ
Isolated yield after preparative HPLC: 82%, α/β 1.9 : 1. The
two anomers proved inseparable by preparative HPLC; purity
> 99% (HPLC); selected NMR-data: δH (300 MHz; CDCl3)
7.20–7.00 (30 H, m, Ph), 5.45 (0.65 H, d, J 3.5, 1α-H), 5.35 (0.65
H, d, J 9.1, α-NH), 5.14 (0.35 H, d, J 8.9, β-NH), 3.31 (1.95
H, s, α-OMe), 3.28 (1.05 H, s, β-OMe), 1.78 (1.95 H, α-Ac), 1.77
(1.05 H, s, β-Ac); δC (75 MHz; CDCl3) 102.9 (22ꢁ, 1Јβ-C), 98.6
(22ꢁ, 1α-C), 98.4 (22ꢀ, 1Јα-C), 97.1 (22ꢀ, 1α-C); m/z (anomeric
mixture) 960.43 ([M ϩ Na]ϩ. C57H63NNaO11 requires m/z,
960.43).
Isolated yield after preparative HPLC: 91% (colorless solid);
mp 132–133 ЊC; purity > 99% (HPLC); observed 1H-NMR data
were identical with literature values:14 δC (75 MHz; CDCl3)
169.7, 166.3, 166.0, 165.4, 165.1, 138.5, 138.1, 133.6, 133.4,
133.33, 133.29, 130.0–127.9 (m), 101.7, 98.5, 80.5, 78.6, 75.0,
74.7, 73.1, 72.5, 72.1, 70.4, 70.1, 68.9, 63.5, 54.9, 52.5, 23.7; m/z
1016.36 ([M ϩ Na]ϩ. C57H55NNaO15 requires m/z, 1016.35).
Methyl 2-acetamido-3,6-di-O-benzyl-2-deoxy-ꢀ-D-gluco-
pyranoside 19
Methyl 2-acetamido-3,6-di-O-benzyl-2-deoxy-4-O-(2,3,4,6-
tetra-O-benzoyl-ꢁ-D-glucopyranosyl)-ꢀ-D-glucopyranoside 23ꢁ
Benzylidene 18 was ring-opened by the method of Debenham
and Toone20 using Et3SiH and BF3ؒEt2O. Isolated yield of 19
after preparative HPLC: 74% (colorless solid), mp 133–135 ЊC
Isolated yield after preparative HPLC: 35%, pure β; purity
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182
2181

View Article Online
ed. M. Fukuda and O. Hindsgaul, IRL Press, Oxford, 1994; (c) K. J.
Yarema and C. R. Bertozzi, Curr. Opin. Chem. Biol., 1998, 2, 49.
4 (a) P. Lerouge, P. Roche, C. Faucher, F. Maillet, G. Truchet,
J. C. Promé and J. Dénarié, Nature, 1990, 344, 781; (b) I. M. Lopez-
Lara, J. D. J. van den Berg, J. E. Thomas-Oartes, J. Glushka, B. J. J.
Lugtenberg and H. P. Spaink, Mol. Microbiol., 1995, 15, 627;
(c) J. Dénarié, F. Debellé and J.-C. Promé, Annu. Rev. Biochem.,
1996, 65, 503.
5 (a) U. Schmid and H. Waldmann, Chem. Eur. J., 1998, 4, 494;
(b) H. Schene and H. Waldmann, Chem. Commun., 1998, 2759;
(c) H. Schene and H. Waldmann, Eur. J. Org. Chem., 1998, 1227;
(d ) H. Schene and H. Waldmann, Synthesis, 1999, 1411.
6 A. Lubineau and B. Drouillat, J. Carbohydr. Chem., 1997, 16, 1179.
7 (a) L. Petersen and K. J. Jensen, J. Org. Chem., 2001, 66, in press;
(b) L. Petersen and K. J. Jensen, presented at the 20th International
Carbohydrate Symposium (ICS), Hamburg, Germany, 2000, August
27–September 1, Poster B-193.
> 98% (HPLC); δH (500 MHz; CDCl3) 8.02–7.78 (8 H, m, Ph),
7.58–7.20 (22 H, m, Ph), 5.70 (1 H, dd, J 9.8, 3Ј-H), 5.58 (1 H,
dd, J 9.8, 4Ј-H), 5.53 (1 H, dd, J 9.8 and 7.7, 2Ј-H), 5.10 (1 H, d,
J 8.5, NH), 4.96 (1 H, d, J 12.4, PhCH2), 4.85 (1 H, d, J 7.7,
1Јα-H), 4.78 (1 H, d, J 11.9, PhCH2), 4.69 (1 H, d, J 3.8, 1α-H),
4.64 (1 H, d, J 11.9, PhCH2), 4.54 (1 H, dd, J 11.9 and 3.0,
6Ј-Ha), 4.41 (1 H, d, J 12.4, PhCH2), 4.27 (1 H, dd, J 11.9 and
5.6, 6Ј-Hb), 4.13 (1 H, ddd, J 10.7, 8.5, and 3.8, 2-H), 4.10 (1 H,
t, J 9.0, 4-H), 3.81 (1 H, ddd, J 9.8, 5.6, and 3.0, 5Ј-H), 3.73
(1 H, dd, J 11.1 and 3.0, 6-Ha), 3.58 (1 H, dd, J 10.7 and 9.0,
3-H), 3.50–3.43 (2 H, m, 5-H and 6-Hb), 3.21 (3 H, s, OMe),
1.77 (3 H, s, OAc); δC (75 MHz; CDCl3) 169.9, 166.3, 165.9,
165.3, 165.1, 138.2, 138.3, 133.6 (2 × C), 133.4, 133.3, 130–
128.3 (m), 127.7, 100.8, 98.6, 77.9, 77.5, 74.5, 73.8, 73.4,
72.6, 72.2, 70.4, 70.0, 67.7, 63.1, 55.4, 52.6, 23.5; m/z 1016.36
([M ϩ Na]ϩ. C57H55NNaO15 requires m/z, 1016.35).
8 Aryl glycosides of salicylic acid are significantly more labile under
acidic hydrolysis than are the corresponding phenyl glycosides. This
has been ascribed to acid catalysis by the free benzoic acid moiety,
see (a) B. Capon, Tetrahedron Lett., 1963, 911; (b) B. Capon, M. C.
Smith, E. Anderson, R. H. Dahm and G. H. Sankey, J. Chem. Soc.
B, 1969, 1038.
Methyl 2-acetamido-6-O-acetyl-3,4-di-O-benzyl-2-deoxy-ꢀ-D-
glucopyranoside 27
9 J. B. Laursen, L. Petersen and K. J. Jensen, Org. Lett., 2001, 3, 687.
10 G. Barany and M. Kempe, in A Practical Guide to Combinatorial
Chemistry, ed. A. W. Czarnik and S. H. DeWitt, American Chemical
Society, Washington, 1997, p. 51.
11 For reviews see (a) P. H. H. Hermkens, H. C. J. Ottenheijm and
D. Rees, Tetrahedron, 1996, 52, 4527; (b) P. H. H. Hermkens, H. C. J.
Ottenheijm and D. Rees, Tetrahedron, 1997, 53, 5643; (c) S. Booth,
P. H. H. Hermkens, H. C. J. Ottenheijm and D. Rees, Tetrahedron,
1998, 54, 15385.
From a determination of loading of acceptor 25, the cleavage
mixture was concentrated, and dissolved in CDCl3: purity
> 99% (HPLC); δH (500 MHz; CDCl3) 7.37–7.28 (10 H, m,
PhCH2), 5.30 (1 H, d, 9.4, NH), 4.87 (1 H, d, J 11.1, PhCH2),
4.85 (1 H, d, J 11.9, PhCH2), 4.66 (1 H, d, J 11.5, PhCH2), 4.64
(1 H, d, J 3.4, 1α-H), 4.60 (1 H, d, J 11.1, PhCH2), 4.34 (1 H,
dd, J 11.9 and 2.1, 6-H), 4.26 [1 H, dt, J 9.4 (2 ×) and 3.4, 2-H],
4.24 (1 H, dd, J 11.9 and 4.3, 6-HЈ), 3.79 (1 H, ddd, J 9.8, 4.3,
and 2.1, 5-H), 3.70 (1 H, dd, J 10.2 and 9.0, 3-H), 3.62 (1 H, dd,
J 9.8 and 9.0, 4-H), 3.34 (3 H, s, OMe), 2.06 (3 H, s, OAc), 1.86
(3 H, s, NAc); δC (75 MHz; CDCl3, selected data) 98.9, 80.6,
78.3, 75.3, 75.2, 69.4, 63.0, 55.3, 52.7, 23.7, 21.1; m/z 480.19 ([M
ϩ Na]ϩ. C25H31NNaO7 requires m/z, 480.19).
12 (a) N. K. Terret, Combinatorial Chemistry, Oxford University Press,
Oxford, 1998; (b) B. A. Bunin, The Combinatorial Index, Academic
Press, San Diego, 1998.
13 Some examples are: (a) S. J. Danishefsky, K. F. McClure,
J. T. Randolph and R. B. Ruggeri, Science, 1993, 260, 1307;
(b) L. Yan, C. M. Taylor, R. Goodnow and D. Kahne, J. Am. Chem.
Soc., 1994, 116, 6953; (c) S. P. Douglas, D. M. Whitfield
and J. J. Krepinsky, J. Am. Chem. Soc., 1995, 117, 2116; (d ) J. A.
Hunt and W. R. Roush, J. Am. Chem. Soc., 1996, 118, 9998;
(e) J. Rademann and R. R. Schmidt, J. Org. Chem., 1997, 62, 3650;
( f ) R. Rodebaugh, S. Joshi, B. Fraser-Reid and H. M. Geysen,
J. Org. Chem., 1997, 62, 5560; (g) S. Mehta and D. Whitfield,
Tetrahedron Lett., 1998, 39, 5907; (h) P. H. Seeberger and S. J.
Danishefsky, Acc. Chem. Res., 1998, 31, 685; (i) R. B. Andrade,
O. J. Plante, L. G. Melean and P. H. Seeberger, Org. Lett., 1999, 1,
1811; (j) F. Roussel, L. Knerr, M. Grathwohl and R. R. Schmidt,
Tetrahedron Lett., 2000, 41, 3043; (k) O. J. Plante, E. R. Palmacci
and P. H. Seeberger, Science, 2001, 291, 1523.
14 J. F. Tolborg and K. J. Jensen, Chem. Commun., 2000, 147.
15 For a recent review, see P. H. Seeberger and W.-C. Haase, Chem.
Rev., 2000, 100, 4349.
16 S. K. Chatterjee and P. Nuhn, Chem. Commun., 1998, 1729.
17 This has been used to generate HF in situ for the removal of silyl
protecting groups: S. Mabic and J.-P. Lepoittevin, Synlett, 1994, 851.
18 R. Caputo, H. Kunz, D. Mastroianni, G. Palumbo, S. Pedatella and
F. Solla, Eur. J. Org. Chem., 1999, 3147.
Methyl 2-acetamido-4-O-acetyl-3,6-di-O-benzyl-2-deoxy-ꢀ-D-
glucopyranoside 28
From a determination of loading of acceptor 26, the cleavage
mixture was concentrated, and dissolved in CDCl3: purity
> 97% (HPLC); δH (300 MHz; CDCl3) 7.37–7.21 (10 H, m, Ph),
5.33 (1 H, J 9.4, NH), 5.14 (1 H, J 10.1 and 9.3, 4-H), 4.74 (1 H,
J 3.6, 1α-H), 4.62 (1 H, J 11.5, PhCH2), 4.57–4.49 (3 H, m,
PhCH2), 4.34 (1 H, ddd, J 10.5, 9.0, and 3.4, 2-H), 3.84 (1 H, dt,
J 10.1 and 4.4, 5-H), 3.73 (1 H, dd, J 10.6 and 9.3, 3-H), 3.54
(2 H, d, J 4.4, 6-H2), 3.38 (3 H, s, OMe), 1.93 (3 H, s, OAc), 1.88
(3 H, s, NAc); δC (75 MHz; CDCl3, selected data) 98.6, 77.8,
77.4, 73.8, 72.6, 71.0, 69.5, 55.4, 51.9; m/z 480.19 ([M ϩ Na]ϩ.
C25H31NNaO7 requires m/z, 480.19).
Acknowledgements
19 K. Umemura, H. Matsuyama, M. Kobayashi and N. Kamigata,
Bull. Chem. Soc. Jpn., 1989, 62, 3026.
20 S. D. Debenham and E. J. Toone, Tetrahedron: Asymmetry, 2000, 11,
385.
We thank the Alfred Benzon Foundation (K. J. J.), Danish
Technical Research Council (K. J. J.), Danish Natural Science
Research Council (K. J. J.), Leo Foundation (K. J. J.) and the
Technical University of Denmark (L. P.) for financial support.
21 R. W. Jeanloz, J. Am. Chem. Soc., 1952, 74, 4597.
22 A. Roy, A. K. Ray, A. Mukherjee and N. Roy, Indian J. Chem.,
Sect. B, 1987, 26, 1165.
23 A. B. Charette and P. Chua, Synlett, 1998, 163.
24 (a) K. J. Jensen, J. Alsina, M. F. Songster, J. Vágner, F. Albericio and
G. Barany, J. Am. Chem. Soc., 1998, 120, 5441; (b) J. Alsina,
K. J. Jensen, F. Albericio and G. Barany, Chem. Eur. J., 1999, 5,
2787; (c) J. Alsina, K. J. Jensen, M. F. Songster, J. Vagner,
F. Albericio and G. Barany, in Solid Phase Organic Syntheses,
ed. A. W. Czarnik, Wiley, New York, 2000, vol. 1, p. 121.
25 Y. Sasaki and D. H. Coy, Peptides (Pergamon), 1987, 8, 119.
26 J. F. Tolborg, L. Petersen, K. J. Jensen, C. Mayer, D. L. Jakeman,
R. A. J. Warren and S. G. Withers, in preparation.
27 W.-S. Kim, S. Hosono, H. Sasai and M. Shibasaki, Heterocycles,
1996, 42, 795.
28 M. L. Falcone-Hindley and J. T. Davis, J. Org. Chem., 1998, 63,
5555.
References
1 Preliminary accounts of this work have been reported: L. Petersen
and K. J. Jensen, presented at the 20th International Carbohydrate
Symposium (ICS), Hamburg, Germany, 2000, August 27–
September 1, Poster B-191.
2 The following abbrevations are used: Trimethylsilyl trifluoro-
methanesulfonate (TMSOTf ), 5-(2-formyl-3,5-dimethoxyphenoxy)-
valeric acid (o-PALdehyde), N-[(1H-benzotriazol-1-yl)(dimethyl-
amino)methylene]-N-methylmethanaminium hexafluorophosphate
N-oxide (HBTU), 1-hydroxybenzotriazole (HOBt), glycosyl donors
derived from dinitrosalicylic acid or its regioisomer (DISAL),
1-methylpyrrolidin-2-one (NMP), polystyrene (PS), poly(ethylene
glycol) polyacrylamide (PEGA).
29 S. S. Rana, J. J. Barlow and K. L. Matta, Carbohydr. Res., 1983, 113,
3 (a) A. Varki, Glycobiology, 1993, 3, 97; (b) Molecular Glycobiology,
257.
2182
J. Chem. Soc., Perkin Trans. 1, 2001, 2175–2182